Equilibrium approach towards water resource management ......water directly intothe river (Kavouridis and Koukouzas, 2008). The situation of severe water shortages, water pollution,

lable at ScienceDirect

Journal of Environmental Management 219 (2018) 56e73

Contents lists avai

Journal of Environmental Management

journal homepage: www.elsevier .com/locate/ jenvman

Research article

Equilibrium approach towards water resource management andpollution control in coal chemical industrial park

Jiuping Xu a, b, *, Shuhua Hou a, c, Heping Xie b, Chengwei Lv a, c, Liming Yao a, b

a Business School, Sichuan University, Chengdu 610064, PR Chinab Institute of New Energy and Low-Carbon Technology, Sichuan University, Chengdu 610064, PR Chinac Uncertainty Decision-Making Laboratory, Sichuan University, Chengdu, 610064, PR China

a r t i c l e i n f o

Article history:Received 14 November 2017Received in revised form16 March 2018Accepted 18 April 2018Available online 3 May 2018

Keywords:Industry ecology (IE)Water and waste load allocationStackelberg-Nash gameConflictsUncertainty

* Corresponding author. Business School, SichuanPR China. Tel.: þ86 28 85418191; fax: þ86 28 854151

E-mail address: [email protected] (J. Xu).

https://doi.org/10.1016/j.jenvman.2018.04.0800301-4797/© 2018 Elsevier Ltd. All rights reserved.

a b s t r a c t

In this study, an integrated water and waste load allocation model is proposed to assist decision makersin better understanding the trade-offs between economic growth, resource utilization, and environ-mental protection of coal chemical industries which characteristically have high water consumption andpollution. In the decision framework, decision makers in a same park, each of whom have different goalsand preferences, work together to seek a collective benefit. Similar to a Stackelberg-Nash game, theproposed approach illuminates the decision making interrelationships and involves in the conflict co-ordination between the park authority and the individual coal chemical company stockholders. In theproposed method, to response to climate change and other uncertainties, a risk assessment tool, Con-ditional Value-at-Risk (CVaR) and uncertainties through reflecting parameters and coefficients usingprobability and fuzzy set theory are integrated in the modeling process. Then a case study from Yuhengcoal chemical park is presented to demonstrate the practicality and efficiency of the optimization model.To reasonable search the potential consequences of different responses to water and waste load allo-cation strategies, a number of scenario results considering environmental uncertainty and decisionmaker' attitudes are examined to explore the tradeoffs between economic development and environ-mental protection and decision makers' objectives. The results are helpful for decision/police makers toadjust current strategies adapting for current changes. Based on the scenario analyses and discussion,some propositions and operational policies are given and sensitive adaptation strategies are presented tosupport the efficient, balanced and sustainable development of coal chemical industrial parks.

© 2018 Elsevier Ltd. All rights reserved.

1. Introduction

With the expansion of coal related industries, these water-intensive projects were constructed predominantly in arid west-ern areas over the duration of the 12th Five-Year Plan (2011e2015)of China which will inevitably trigger a serious water crisis andexacerbate existing water scarcity problems (Li and Hu, 2017). Coalchemical industries are not only extremely water-intensive but alsocharacterized by significant pollutionwith a large amounts of wastewater directly into the river (Kavouridis and Koukouzas, 2008). Thesituation of severe water shortages, water pollution, and

University, Chengdu 610064,43.

environmental degradation in coal chemical industrial areas hasbecome increasingly serious and caused a major bottleneck inconstraining sustainable economic and environmental develop-ment (Xie et al., 2010). Motivated by the conflicts between energyeconomy development and water resources and environmentalprotection, this paper aiming to improve the water utilization ef-ficiency and minimize the industry's environmental impact andstresses the importance of regulated water resource allocation andpollution control strategy.

Nowadays, access to clean, affordable, reliable energy has been acornerstone of the world's increasing prosperity and economicgrowth (Chu and Majumdar, 2012). With the development of cleancoal technology, modern coal chemical industry has rapidlydeveloped which is mainly involved with the production of cleanenergy and substitutes from diesel oil, gasoline, etc (China'sNational Energy Administration, 2015). It has been prioritized in

mailto:[email protected]://crossmark.crossref.org/dialog/?doi=10.1016/j.jenvman.2018.04.080&domain=pdfwww.sciencedirect.com/science/journal/03014797http://www.elsevier.com/locate/jenvmanhttps://doi.org/10.1016/j.jenvman.2018.04.080https://doi.org/10.1016/j.jenvman.2018.04.080https://doi.org/10.1016/j.jenvman.2018.04.080

J. Xu et al. / Journal of Environmental Management 219 (2018) 56e73 57

recent years as it is expected to play a critical role in the sustainabledevelopment of energy resources (Xie et al., 2010; Man et al., 2017).The water resource scarcity and pollution problem have restrictedthe development of the coal chemical industry due to the reversedistribution of coal and water which is a prominent feature ofChina's resource endowments (Rubiocastro et al., 2010; Han et al.,2015). To battle with high water consumption and pollution thataccompanies rapid industrial growth, most optimization studieshave been focused on optimizing water exchange network or cleantechnology (Xiong et al., 2017). However, the associated problemsof extensive energy use, water resource inefficiency, and heavypollution have not been integrated to solve effectively (Tanakaet al., 2017; Shen et al., 2017).

With their increased resource efficiency and cleaner, moreenvironmentally friendly production processes, coal-based energyparks have been seen as an efficient method to improve coalchemical industry sustainability (Lin and Long, 2015; Yune et al.,2016). Evidence has shown that encouraging a symbiotic rela-tionship between the coal chemical plants in the same industrialpark can significantly contribute to the sustainable development ofthe industrial activities (Boix et al., 2015). Industrial parks, whichare established based on IE, have a number of independent com-panies and have the aim of concentrating economic developmentand the associated environmental pollution into a geographicallyconfined area Ren et al. (2016). Therefore, these eco-industrialparks (EIPs) present both opportunities and challenges for envi-ronmentally sustainable industrial development Shi et al. (2010).While the industrial concentration in the parks makes it possible toachieve the proposed Sustainable Development Goals (SDG) Griggset al. (2013); UN (2016) and gain economic and environmentalbenefits, this symbiosis between the industrial activities could leadto pollution and resource depletions that exceed local environ-mental carrying capacities. By applying a new paradigm of princi-ples and system tools, IE provide concrete guidance for industrialpark decision support activities Edgington (1995); Pauliuk et al.(2017). IE, as a solution-driven approach, integrates the environ-mental concerns into the production and resource utilizationstrategies with research having suggested a variety of practicalinitiatives. Comprised with those study which focused on certainpoint of resource or environmental problem, there has been lessattention paid to developing a systematic, comprehensive meth-odology for reducing pollution, water resource saving, andincreasing economic effectiveness (Zhang et al., 2015; Xinchunet al., 2017).

Water use and environmental protection problems have oftenbeen resolved using water allocation or quality controlsuperstructure-based model, where the water is optimallydistributed, treated and discharged in a region (Lovelady and El-Halwagi, 2009; Xie et al., 2017). Integrating water resource man-agement and environmental protection through water and wasteload allocations has also been studied for water resource andenvironment management in a river basin (Wang and Huang, 2011;Zeng et al., 2017a). Few total pollution control management tech-niques have been integrated into water environmental manage-ment systems of a industrial park. Therefore, to address thisresearch gap, in this paper, total pollution, quality control, andwater resource efficient allocations are integrated in a systematicmodel, in which the decision framework incorporates both themanagement objectives of the individual companies and thebalanced development of the park as a whole, and accounts for theconflicting objectives of the participants who prioritize individualclean production based on IE and the park authority who prefers tofocus on the overall economic growth while maintaining orimproving the park's environmental performance (Leong et al.,

2017).The park authority and the coal chemical company stake-

holders who the have conflicting objectives and preferences asdecision makers are participated in the water resource and envi-ronmental management of the park. Decisions are made sequen-tially from the upper level to the lower level, with each leveldecision makers having different powers of control over themanagement objectives and decisions. Therefore, this decisionsituation fits with a decentralized decision system in which oneleader and several followers of equal status are involved (Liu,1998). A powerful tool dealing with this decision systems is theso-called bi-level programming (Wen and Hsu, 1991;Anandalingam and Apprey, 1991). And the conflict betweeneconomically efficient growth and environmental protection, orthe tension between the economic utility of the natural resourcesand their ecological utility in the natural environment, has notbeen completely solved under EIP development as excessive waterconsumption and high emissions necessitate a trade-off betweeneconomic development and environmental sustainability. In thisproblem, to coordinate the relationship between the economy andthe environment, a multi-objective method which involve morethan one objective function is applied to model the optimal waterand waste load allocations for the entire industrial park (Xu et al.,2011, 2016b; Leong et al., 2017). The set of solutions gained frommulti-objective optimization model define the best tradeoff be-tween competing objectives that is more in line with the problemsbeing studied. As the external uncertainties in this integratedwater resource and waste load allocation optimization modelcould result in different water and waste allocation strategies, arisk evaluator of the conditional value at risk (CVaR) which areemployed to assess the impacts of degrees of the preference ofdecision makers on the tradeoff between system benefits andexpected economic losses (Wang et al., 2017). And probability andfuzzy set theory are used to deal with uncertainty parameters andvariables. Scenarios are then examined to evaluate the un-certainties related external environmental change, and the deci-sion maker' attitudes, and to develop sensitive adaptationstrategies to mitigate environmental damage and adapt to possibleclimate conditions (Wu et al., 2017). Motivated by the contradic-tions between economic development, environmental protection,and climate-change mitigation in the coal chemical industrialpark, this paper develops an equilibrium approach for efficientenergy conservation, water utilization and pollution control basedon IE (Tang et al., 2015; Niu et al., 2017).

2. Key problem statement

To improve the efficiency of energy conservation, water utili-zation, and ensure environmental protection, a decision makingframework, as shown in Fig. 1, is developed aiming at collectivebenefits achievement with equilibrium water and waste load allo-cation strategies.

2.1. Coordinating economic and environmental conflicts based on IE

Significant environmental problems rose with rapid develop-ment of coal chemical industrial, particularly as the worseningenvironmental conditions have resulted in major challenges toeconomic growth (Chu and Majumdar, 2012). By incorporatingmulti-objective procedures into the decision making process, thepark decision maker attempts to balance the economic and envi-ronmental goals so that the proposed model is more realistic andresults in an ecological solution to coal chemical industrial opera-tions. As a multi-objective optimization problem is able to search

Fig. 1. Decision framework for water utilization and pollution control.

J. Xu et al. / Journal of Environmental Management 219 (2018) 56e7358

for trade-off solutions under various scenarios, a trade-off curveknown as the Pareto front is generated, which provides a greaternumber of alternative strategies for decision or policy makers. Asolution-driven approach is proposed through the design ofappropriate resource utilization and production strategies to helpminimize environmental damage while improving economic per-formance (Karn and Bauer, 2001), thereby resolving the conflictsbetween economic growth and environmental protection (Alper,1992).

2.2. Seeking eco-industrial collective benefits and conflict trade-offs

In this problem, a group of decision-makers with conflictinginterests develop an eco-industrial park under an IE framework(Shi et al., 2010). The park authority and private stakeholders worktogether to gain collective benefits rather than only optimizingeach individual performance (Liu et al., 2015). In this paper, thedifferent hierarchical decision making levels of the park authority/policy maker and the company stakeholders, each of which have

Indexi Index for coal chemical companies located in the same industrial parkj Index for coal chemical projectst Index for target control pollutantsDecision variables

YfiFreshwater allocated to company i

Ywi Waste load for pollutant t allocated to company i

xfijAllowable water consumption for the jth coal chemical projectproduction at company i

xwij Allowable water waste discharge of the jth coal chemical productproduction at company i

Certain Parameterssci Unit removed costs for the pollutantsEðtÞ Current emissions of pollutant tt Pollution tax per unit of pollutant tRðtÞ Emissions reduction rate for pollutant tURi Unit costs for wastewater reuse in company iCsðtÞ Standard effluent concentration of pollutant tPCij Production capacity for project j at company iPDij Minimum yield demand for project j at company ibmax Maximum economic return per unit of water consumptionbi Industrial output value per unit of water consumption at company iri Pollutant removal rate through the sewage treatment facilities

min


conflicting objectives, are integrated in the decision making systemframework. The park authority is responsible for improving theeconomic performance of the park while minimizing the environ-mental impacts (Tian et al., 2014), and has the higher authority formanaging the environmental, water and energy use issues, all ofwhich can influence the performance of the individual stake-holders through the implementation of policy or incentive mech-anisms. The participating companies are the followers who need toabide by the park authority's decisions but are able to individuallyembed IE based clean production and pollution prevention strate-gies in their industrial activities. The followers react to the leader'sstrategy and seek to develop collective optimal strategies. The de-cision making process between the park authority and the indi-vidual companies is therefore similar to a Stackelberg game(Stackelberg and Peacock, 1953; Stackelberg et al., 2011). It em-phasizes trade-offs between two levels of objectives with theconflict between these stakeholders finally arriving at aStackelberg-Nash equilibrium (Nash, 1951). The solution to the bi-level water and waste load allocation optimization problem is theStackelberg-Nash equilibrium (Liu,1998) that can achieve collectivebenefits for the entire eco-industrial park.

WDi Minimum water resource demand to maintain production at company i

v Equivalent value for the different pollutants that measures theenvironmental harm

Uncertain Parametersemi Water recycling rate inside the company iegi Wastewater discharge coefficient at company igblossi

Water transfer loss ratio from supply centre to company i

gueij Unit operating costs for the jth coal chemical project at company igucij Unit water consumption for the jth chemical project production atcompany i

~Q Uncertainty of available water resources from the external environment

Risk attitude parametersa Park authority's risk-aversion degreeε Park authority's tolerance level for economic efficiency loss

2.3. Adaptation strategies for environmental changes

A systematic decision making framework is designed to supportefficient water use and pollution control and respond to environ-mental changes with the aim of achieving sustainable develop-ment. As eco-industrial park water supplies are usually taken fromrivers, groundwater or recycled water resources, with the capacityof the water bodies depending on many meteorological and hy-drological factors, there are many uncertain features (Walther et al.,2002), such as changes in rainfall, all of which can impact thequality and quantity of water resources available to the industrialecosystem (Xu et al., 2013, 2015b). To cope with these extensivewater resource and water environment system uncertainties,probability and fuzzy set theory are used to deal with the proba-bilistic and discrete interval uncertainties in water and waste loadallocation problems. Available water fluctuations are the mainwater resource allocation uncertainties, as these directly affect in-dustrial production, thereby causing economic efficiency loss risks(Kong et al., 2017). This is especially evident in the coal chemicalindustries which have high water and energy consumption andhigh pollution (Hu et al., 2016; Zhuang et al., 2017). In this paper,the conditional value at risk (CVaR) is used as the risk evaluator tomeasure the economic losses across the entire coal chemical park invarious scenarios to assess the adverse impact of environmentalchanges on the delivery of ecosystem services. Uncertain parame-ters, such as water flow can be expressed as a probability distri-bution; however, as the other uncertain factors cannot beexpressed as probability density functions (PDFs) due to the lack ofavailable monitoring data, fuzzy sets can be used to identifyeffective water resource and pollution management alternatives(Poff et al., 2016; Zeng et al., 2017b). Therefore, based on the sus-tainability insights gained from IE, to respond to the environmentalchanges, this paper develops adaptation strategies through variousscenarios to support improvements in the economic and environ-mental performances of both the individual park companies andthe park as a whole.

3. Modeling

The mathematical description of the above conceptual modeland an efficient solution approach is presented in this section.

3.1. Assumptions

Before establishing the model in this paper, the following modelassumptions are outlined:

1. Water and waste load allocation period is set as one year. Beforethe next plan cycle begins, a new water resource and environ-mental management plan is developed under the scenario atthat time.

2. The lower-level decision makers, coal chemical companies, areable to receive water from only two sources (i.e., fresh and re-generated water).

3. The operating costs of a coal chemical project is assumedincluding capital costs, energy costs, and water consumptioncosts.

4. The main pollutants, chemical oxygen demand (COD) andammonia nitrogen (NH3-N), are as the control factors to eval-uate the environmental damage in this paper.

3.2. Model development

This section describes the development of the proposed bi-levelmulti-objective mathematical model.

3.2.1. Coordinating overall economic and environmental parkperformance

To determine the economically and environmentally optimalstatus for the industrial ecosystem, the park authority has to tradeoff efficient economic growth and environmental protection. Basedon IE theory, the objectives of the decision maker, therefore, are to


minimize the economic efficiency loss risks while minimizing totalpollution to reduce environmental damage.

Minimizing total pollutionPollutants emitted from the coal conservation process can cause

direct or indirect environmental damage. While some treatmenttechnologies can reduce the emissions, they are unable to radicallycontrol environmental pollution damage. To control environmentalpollution, previous studies have sought to implement total waterpollutant control programs through the setting of effluent stan-dards (Jia et al., 2016).

However, it is more efficient to develop integrated controlstrategies rather than develop separate strategies for each indi-vidual pollutant. As there are many pollutants produced as a resultof the coal chemical production process, the park authority needs toimplement controls to minimize the total environmental impact.

F1 ¼XTt¼1

Xmi¼1

Ywi ðtÞvðtÞ (1)

where Ywi ðtÞ is the waste load of pollutant t allocated to company i,and vðtÞ refers to the equivalent value of the different pollutantsbased on their degree of harm to the environment and technicaltreatment economics (People, 2008).

Economic efficiency loss risk aversion objectiveAs the high water consumption needs of the coal chemical in-

dustries can aggravate fresh water scarcity and further affect eco-nomic safety, CVaR is developed to coherently measure theexpected economic losses as the actual losses exceed the Value atRisk (VaR) threshold (Moghaddam et al., 2013). Based on knownprobability distributions of random variables, CVaR is designed tomeasure and quantify the level of the likelihood of incurring lossescaused by water resource shortage. Since the scope of risk assessedis limited and the tail end of the distribution of loss is not typicallyassessed, if losses are incurred, the amount of the losses will besubstantial in value.

The objective is to estimate the economic efficiency for a pre-

scribed probability level. Let L½ðYfi ; ~QÞ� be the risk-averse economicefficiencymeasure for the park authority that is associated with the

water allocation volume Yfi decision variable at the ith coal chem-ical company. The park economic efficiency is defined as the coalchemical production economic returns divided by the maximumpotential economic returns under uncertain water resource flows.

L�Yfi ;

~Q�¼ 1�

Pmi¼1

bi

0@1� E24gglossi

351AYfibmax

~Q þPm Pn

j¼1liE

"ebi#xfij

!

where bi is the industrial output value per water unit at company iand bmax is the maximum economic return per unit of water con-sumption in the coal chemical industry. The CVaR value at a specificconfidence level a3½0;1� can therefore be calculated as follows;

CVaRa ¼ EnLh�

Yfi ;~Q�i��J�Yfi ; ~Q� � ao

where the cumulative distribution function of loss at decisionvector x is Jðx;xÞ ¼ Prff ðx;xÞ � ag.

To ensure a mathematically well-behaved risk measure, CVaRacan be expressed using the following linear minimization formula,which minimizes the following linear function over both a and thedecision variable x that CVaRa has been shown to be equivalent tothe following model (Rockafellar and Uryasev, 2010):

CVaRa¼minFaðx;zaÞ¼zaþ1

1�aXMs¼1

Psh�Lh�

Yfi ;~Q�i

þzaiþ

(2)

where zaðYfi Þ is the VaR value that represents the maximum loss ata specific confidence level a, a represents the park authority risk-aversion degree, the smaller the a, the more risk-averse; and 1-adenotes the accumulated probability that the water use economic

efficiency losses are lower than zaðYfi Þ.

(1) Water resource availability constraints

Xmi¼1

Yfi � ~Q þXmi¼1

Xnj¼1

liE

"ebi#xfij (3)

where ~Q is the uncertainty of availablewater resource from the river

flow andPm

i¼1Pn

j¼1E½ebi�xfij is the wastewater produced in the park,in which ebi is the wastewater discharge coefficient, which becauseof the influence of various objective factors can not be exactlyvalued; however it can be estimated and as using typical triangularfuzzy numbers, with the most likely value being in a relatively

smaller range. Based on fuzzy set theory, ebi can be described as ebi ¼ðki1; ki2; ki3Þ with a most possible value of ki2 and the lower andupper bounds, ki1 and ki3. However, this parameter can not bedirectly calculated, in this paper, the expected value operatormethod for fuzzy numbers proposed by (Xu and Zhou, 2012) is used.To efficiently utilize the limited fresh water resources and reducethe wastewater discharges that can cause water environmentpollution damage, as much water as possible should be recycled;therefore, l represents thewastewater reuse rate,whichdepends onthe overall plan for reducing water consumption across the park.

(2) Water consumption constraint

The supplied water should be larger than the minimized freshwater demand:

WDmini �0@1� E

24gglossi351AYfi

where WDmini is the minimum water required to maintain pro-duction at company i. gi is the water loss coefficient due to evap-oration, leakage, drainage piping and other losses in the coal

conservation and effluent processes, and E½gglossi � is the expectedvalue of the water transfer loss ratio from the supply centre tocompany i.

(3) Pollution emissions reduction requirements

To balance the relationship between economic developmentand environmental protection, total pollutant emissions should notexceed the load capacity of the external environment; therefore,the park authority must ensure a low level of total pollutantdischarge;

Xmi¼1

Ywi � ð1� RðtÞÞEðtÞ; c t

where EðtÞ is the current pollutant emissions, with RðtÞ being theplanned pollutant emissions reductions over the year.


3.2.2. Cost-effective clean production for the individual parkcompanies

Under the ecological policy requirements, each coal chemicalcompany must implement environmental protection facilitiesconstruction to reduce environmental damage (Service, 2017).Therefore, the effect on production costs of the environmental taxpayments (Xie et al., 2010), and any increase in coal prices cannotbe ignored. Minimizing clean production costs must therefore ac-count for unit operation costs, wastewater reuse costs, and thepollution tax;

TCi ¼Xnj¼1

E�gueij� xfij

E�gucij�þ URili

Xnj¼1

E

"ebi#xfij þ

XTt¼1

Xnj¼1

vðtÞ½

� ðsciriðtÞ þ tðtÞð1� riðtÞÞ�xwij ðtÞ (4)

whereueij is the unit energy costs for the production of a unit of the j

chemical product at company i, E½gueij� is the expected value of thetrapezoidal fuzzy number gueij, E½gucij� is the expected water con-sumption per unit for the j chemical production project at companyi. URi is the wastewater recycling costs and riðtÞ is the reusedwastewater percentage. For the total pollutants produced byall parkindustrial activities, ri of them are removed in the sewage treatmentfacilities with an expected unit removed cost of sci for coal chemicalplant i. Let t be the unit pollutant tax per pollutant t unit that coal

chemical company i must payPT

t¼1Pn

j¼1vðtÞ½tðtÞð1� riðtÞÞ�xwij ðtÞ forthe damage to the environment.

Each plant in the park develops clean production plans tominimize total costs under the water and pollutant discharge rightsallocated by the park authority. Then, the park coal chemicalcompanies decide on individual water and waste load allocationstrategies to support clean-efficient production, which are then fedback to the upper level decision system. Therefore, as mentioned,the park authority' goals are impacted by Stackelberg game theory(Xu et al., 2016a; Yao et al., 2016).

(1) Available water resource constraint

The total water resources allocated to all park coal chemicalprojects cannot exceed the available water rights issued by the parkauthority.

Xnj¼1

xfij ��1� E

�egi��Yfi(2) Pollutant emissions constraints

To reduce the water environment damage, total pollution con-trol is the priority of IE target. As pollution discharge rights areassigned to each park company, the waste load allocations clearlyidentify the pollutant discharges permitted for each productionproject.

Xnj¼1

xwij � Ywi

(3) Water quality constraints

As an important tool in water environment management, waterquality controls can provide a reasonable guide to the sewage dis-charged into the external water environment. The more controlimposed on water quality, the lower the risk of unforeseen

pollution emergencies.

Pnj¼1

ð1� riðtÞÞxwij ðtÞ

ð1� miÞPnj¼1

E

"ebi#xfij

� CsðtÞ

The total pollutants discharged to water environment that notincluded the removed pollutants by treatment facilities arePn

j¼1ð1� riðtÞÞxwij ðtÞ. And the wastewater could be pumped into theriver system is ð1� EmiÞ

Pnj¼1E½ebi�xfij with a recycling rate mi within

the company. Where CsðtÞ is the emission standard for pollutant tbased on the water quality standards.

(4) Product yield constraints

Each coal chemical project produces products to meet theminimum market demand under the production capacity con-straints imposed by the water resource,

PDminij �xfij

E�gucij� � PCij

where ucij is the water demand per unit of coal chemical product iproduced at company i, and PCij is the production capacity ofproduct j at company i.

3.2.3. Global modelGuided by the IE principles, a bi-level multi-objective water and

waste load allocation model is established, as the global modelshowed in Eq. (5), to achieve efficient water utilization and reduceenvironmental damage (Jia et al., 2016). Each individual coalchemical company can achieve cleaner production and pollutionreductions through the integration of IE tools and approaches. Incoal chemical industrial parks where all participants share thesame eco and industrial systems, the park authority and the indi-vidual stakeholders work collectively in an integrated system toachieve higher benefits than can be achieved individually (Graftonet al., 2013). As this model is focused on an eco-industrial park, itnot only aims to coordinate the conflicts between economicdevelopment and environmental protection, but also seeks toresolve the conflicting objectives between the park authority on theupper level and the company stakeholders on the lower level. Asthe leader, through the water resource and pollutant emissionrights allocations, the park authority seeks to minimize waterpollution emissions with minimum economic efficiency loss risk,and based on these water and waste load allocations, the individualpark stockholders develop IE based production plans to improvetheir individual water use efficiency and minimize costs throughthe reallocation of the water resources and waste loads to differentcoal chemical projects. The park authority and park companiesoptimize their objective functions without considering the otherpartys objectives; however, the decisions of each party affect boththe achievement of these individual objectives and their decisionspace. The Stackelberg game between the two levels of decisionmakers allows for a trade-off between the ecological and industrialsystems (Boix et al., 2015). Through an energy and resource utili-zation and environmental protection synthesis in the decisionmaking framework, the proposed model provides a promisingstrategy to reconcile conflicts between economic growth andenvironmental protection and objective conflicts of decisionmakers (Edgington, 1995).

minF1 ¼XTt¼1

Xmi¼1

Ywi ðtÞvðtÞ

minF2 ¼ za þ1

1� aXMs¼1

PshL�Yfi ;

~Q�� za

iþ

8>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>:

s:t:

L�Yfi ;

~Q�¼ 1�

Pmi¼1

bi

0@1� E24gglossi

351AYfibmax

0@~Q þ lXmi¼1

Xnj¼1

E

"ebi#xfij

1AXmi¼1

Yfi � ~Q þ lXmi¼1

Xnj¼1

bixfij

WDmini �0@1� E

24gglossi351AYfi

Xmi¼1

Ywi � ð1� RðtÞÞEðtÞ; c t

Yfi � 0 Ywi ðtÞ � 0

minTCi ¼Xnj¼1

E�gueij� xfij

E�gucij�þ URili

Xnj¼1

E

"ebi#xfij

þXTt¼1

Xnj¼1

vðtÞ½ðsciriðtÞ þ tðtÞð1� riðtÞÞ�xwij ðtÞ

8>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>:

s:t:

Xnj¼1

xfij �0@1� E

24gglossi351AYfi

Xnj¼1

xwij � Ywi

Pnj¼1

ð1� riðtÞÞxwij ðtÞ

ð1� miÞPnj¼1

E

"ebi#xfij

� CsðtÞ

Dminij �xfij

E�gucij� � PCij

xfij � 0 xfij � 0

(5)


The solution-driven IE approach coordinates the conflicts be-tween economic performance and environmental sustainability byaddressing the conflicting objectives of the decision makers on thetwo levels (Liu et al., 2015), and providing water resource andenvironmental management system equilibrium strategies (Xuet al., 2015a). The proposed model mathematically deals withboth the bi-level conflicts and the multi-objectives to suggestsustainability solutions for any future water environment changesthrough the application of CVaR risk measurement technology andprobability and fuzzy set theory to tackle the uncertain parameters.Sensitive adaptation strategies are then developed through variousscenarios based on the solution-driven approach presented in thissection that respond to the environmental changes impacting theindustrial ecosystem. Themodel can be applied to similar situationsby adjusting the policy control parameters or by changing, addingor removing the park authority objective functions for the con-straints to address complex systems integration and identifyeffective solutions to sustainability challenges.

3.3. Model solution approach

The water and waste load allocation programming model with asingle decision maker, the park authority, on the upper level, andmultiple individual coal chemical company decision makers on thelower level is shown in Eq. (5). In the bi-level model, the waterresource and waste load decision variables are divided into two

subsets, one of which is controlled by the leader (Yfi and Ywi ðtÞ) and

the other of which is controlled by the followers (xfij and xfij). The

followers resolve the inner problem to optimize total costs (Eq. (4)).

The leader then examines each followers' feasible choice xfij and xfij,

with the feasible region of the upper level optimization problembeing determined by the leader's own constraints as well as theinner optimization problem (Sinha and Sinha, 2007). As thisproblem is a nonconvex problem and therefore complicated tosolve (Saharidis and Ierapetritou, 2009), in this paper, a Karush-Kuhn-Tucker (KKT) transformation method, which is a decision-making process from a lower to higher level, is applied. The KKToptimal conditions are incorporated into the lower level pro-gramming cost minimization objective (Eq. (4)) and the resourceand production capacity constraints. By employing the Lagrangemultipliers on the constraints, as shown in SupplementaryMaterial, the problem is transformed into an equivalent (one-level) mathematical program under appropriate constraints.

On the upper level, the park authority must simultaneouslyhandle multiple conflicting goals to achieve economic benefits andminimize environmental damage. As these economic and envi-ronmental goals are expressed in different measurement units, theintegration of the various goal-deviations in their original formshas no practical significance in Eqs. (1) and (2). To overcome thesedifferences, this paper implements a ε-constraint method for pro-ducing the Pareto optimal solutions for the multi-objective plan-ning model. The multi-objective optimization problem is thentransformed to an IE based single-objective optimization problem(Qi and Jiang, 1997) with the objective being expressed in Eq. (1)and the additional constraints being expressed in Eq. (2) using aconstraint vector ε. A bi-level model is very similar to a multi-objective problem as they both have conflicting objectives. In thetransformed model, optimum computing solutions can achieveequilibrium for the various decision makers and trade-off theconflicting objectives. Therefore, in this paper, ε-constraint methodand the KKT method are integrated to solve the proposed model.


4. Case study

In this section, the Yuheng coal chemical industry park is takenas a case study to demonstrate that practicality of the proposedmodel.

4.1. Study area

The Yuheng coal chemical industry park is located in Yulin Cityon the Northern Shaanxi Loess Plateau and Mu Us Desert borderarea in China. While the area has large reserves of good quality coal,it has poor water resources, with the per capita water resourcesbeing only 865m3 which is far below the provincial and the na-tional averages. As one of the “two parks and six districts” in Yulin'senergy development plan, the Yuheng coal chemical industrial parkhas a total planned area of 48.5 km2, as shown in Fig. 2. Five largestate-owned energy and chemical companies with varying projectsare located in the Yuheng coal chemical industrial park. The mainwater supply comes from the Wang Gedu reservoir located up-stream of the Wuding river. The Wang Gedu reservoir watertransport project report claimed that the reservoir can guarantee155 million m3 of water to the park industrial area every year.However, according to the Overall plan for the Shaanxi Yuheng CoalChemical Industry Park (2010e2030), the current water con-sumption demand by the park industries is higher than supply;therefore, it has become urgent to address this water scarcityproblem.With the accelerated pace of industrial construction in thepark, besides the water shortage problems, there are also issueswith poor quality water. Therefore, as this conflict betweencontinued industrial development, limited water supplies, andenvironmental pollution could severely restrict future park

Fig. 2. Geographic location of Yuhen

development, to ensure the stable, clean, sustainable developmentof the industrial park, it is necessary to devise a strategy that effi-ciently utilizes the water resources and effectively controlspollution.

4.2. Data collection

The data were taken from the “Report on the EnvironmentalImpact Assessment of the Development Plan for the Yuheng In-dustrial Zone” (Yulin High-tech Industrial Development ZoneManagement Committee, 2017) and regional public informationand statistical yearbooks. From the “Overall plan for the ShaanxiYuheng Coal Chemical Industry Park (2010e2030)” and the “Yulinhigh-tech Zone 13th Five-Year Plan”, sustainable developmentforecasts for the minimum freshwater demand for each companyand the production capacities for each project were determined, asshown in Table 1.

The industrial park is in a dry region that has little rain and highevaporation. The surface water resources are small and the spatialand temporal distribution of the surface water runoff is uneven,which means that the water supply in this area is limited and un-certain. To define the stochastic parameters for the freshwater,scenarios and probability, mass functions based on Gaussianquadrature are used in the following steps (Miller and Rice, 1983);(1) based on historical data, the mean and coefficient of variation

(CV) for the ~Q are estimated; (2) it is assumed that ~Q complies witha normal distribution, for which the expected value and standarddeviation (SD) scenarios are simulated and as many moments aspossible from the original distribution preserved. For the totalfreshwater resources from the external environment, four sce-narios are developed for the three mass normal distribution

g coal chemical industrial park.

Table 1Basic data of the five coal chemical companies in Yuheng coal chemical park.

Company Abbreviation bi blossi WDmini

Project PCij UCij

(i) (RMB/104 t) (%) (104 t/a) (j) (104 t) (t/t)

China Coal Shaanxi Yulin Energy Chemical Co., Ltd. Yanchang Petroleum 565.27 0.08 2739.92 Coal-to-Methanol(CTM) 225 12.94Methanol-to-Olefin(MTO) 70 11.2Methanol-to-Propylene (MTP) 70 11.2

Shaanxi Yanchang Petroleum (Group) Co., Ltd. Yanchang Petroleum 326.56 0.12 245.56 Coal-to-Methanol(CTM) 100 12.94Coal based Acetic Acid(HAC) 160 11.2Coal-to-Oil (Synthetic oil) (CTL) 25 7.12

Shaanxi Yulin Excelle Coal Chemical Co., Ltd Yulin Excelle 40.282 0.14 350.7 Coal-to-Methanol(CTM) 300 12.94Coal derived Aromatics(CTA) 100 279Coal-based PVC 60 10.5

Shaanxi Future Energy Chemical Corp Future Energy 82.74 0.09 1147.3 Coal to Oil (indirect liquefaction) (CTL) 400 7.12

Table 2Uncertain water resource parameter.

Scenario Available freshwater resources ~Q (108 m3) Probability p

Low Medium High

(CV1 ¼ 0:12, s1 ¼ 0:1528) (CV2 ¼ 0:18, s1 ¼ 0:2831) (CV3 ¼ 0:25, s3 ¼ 0:3182)S1 1.6294 1.9337 2.0157 0.045876S2 1.3862 1.4829 1.5089 0.454124S3 1.1595 1.0628 1.0367 0.454124S4 0.9163 0.6119 0.5300 0.045876


functions with equal expected values and different SDs. From the“Research Report for the Wang Gedu reservoir project in Shaanxi,Yulin”, 0.12, 0.18, 0.25 are chosen as the CV values, with an expectedvalue of 127.28�106m3 for ~Q. Accordingly, these three distributionscorrespond to a low, medium, and high level of variation. Thespecific scenarios and corresponding probabilities are shown inTable 2.

4.3. Results and discussion

In this section, optimization results under different decisionmaker attitudes for the confidence levels (1� a) and the economicefficiency lose tolerance level (ε), and environmental conditions(i.e. water flow variation coefficient, CV) are given.

4.3.1. Scenario results

4.3.1.1. Scenario 1: CV¼ 0.12, ε¼ 0.40, and a changing 1-a. In thisscenario, the risk-based optimization model is run under a stablewater resources environment condition in which the water re-

sources ~Q has a variation level of CV¼ 0:12. The park authorityattitude towards the CVaR constraint is set at 0.40 (ε ¼ 0.40),indicating that the risk averse decision maker has a relatively hightolerance for economic efficiency losses in the park; that is, thehighest level economic efficiency loss of 40%. By changing theconfidence levels (1-a) in Eq. (2), a group of results were derived, asshown in Table 3. When 1-a¼ 0.99, the total pollution denoted bythe comprehensive pollution equivalent value expressed in Eq. (1)is 5966.33 tonnes $ yr�1. It can be seen that the total pollution valueincreases noticeablywith an increase in 1-a. From the results in thisscenario, the total pollution produced by the whole park can becontrolled within 5966.33 tonnes when the whole park economicefficiency loss is set at 40% or less under a low water resourcevariation.

4.3.1.2. Scenario 2: CV¼ 0.12, 1-a¼ 0.99, ε is changing. In this sce-nario, when the water resource is at a low variation level, the parkauthority has a 99% confidence interval. By changing the economic

efficiency loss tolerance level, a group of results were derived thatdemonstrate the different decision maker attitudes towards theeconomic efficiency loss tolerance and the influence on the allo-cation strategies. From Table 4, it can be seen that total pollutionincreases if the decision maker expects a lower economic efficiencyloss, which reflects the conflict between economic developmentand environmental protection. In the current resource conversionefficiency situation, to achieve greater economic returns, the parkenvironmental damage cannot be controlled at a lower level. Underthe low water resource variation environmental condition, theeconomic efficiency losses can be controlled at 0.1 with themaximum pollution being 8611.78 tonnes at a fixed confidencelevel of 1-a¼ 0.99. It is worth noting that the allocation results forthe China Coal company are not sensitive to the changes inÅ;therefore, the other three companies can gain a larger share of thewater and waste load allocations than when ε decrease to 0.20 to0.40 with a step length of 0.05.

4.3.1.3. Scenario 3: CV¼ 0.18, ε ¼ 0:40, 1-a is changing. In thisscenario, the attitude towards the CVaR constraint is set at 0.40 (ε ¼0.40), and the risk-based bi-level model is run under a moderateenvironmental condition with a medium variation in the waterresource supply; that is CV¼ 0:18. By changing the confidence level(1-a), a group of results are determined, as shown in Table 5. In thisscenario, if ε is set at 0.40, total pollution can be controlled to be-tween 5772.15 and 6833.13 tonnes. As the results show, as theconfidence level (1-a) increases from 0.80 to 0.99, the total pollu-tion continues to grow, which indicates that a larger degree of riskaversion results in greater pollution, which obstructs the ecologicalobjectives of the coal chemical industry. From the allocations toeach company, it can be seen that Yanchang Petroleum and theFuture Energy companies are allocated insensitive water and wasteloads from the park authority, and that the allocations to the YulinExcelle company increase when 1-a increases, indicating that thepark authority intends to achieve a more accurate evaluation of theeconomic efficiency losses.

4.3.1.4. Scenario 4: CV¼ 0.18, 1-a¼ 0.99, ε is changing. Under a

Table 3Optimal results under various confidence levels with CV¼ 0.12 and ε¼ 0.40.

Scenario indicators Total pollution(tonne)

Allocated toCompanies

Water allocations(104m3)

Allocated loads Allocated toprojects

Allocation results Totalcosts

COD(tonne)

NH3-N(tonne)

Water(104m3)

COD load(tonne)

NH3-N load(tonne)

(108

RMB)

(CV¼ 0.12, ε¼ 0.40, 1-a¼ 0.99)

5966.33 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97

Yanchang 1859.31 680.23 67.08 CTM 647.00 268.98 26.53 37.14Petroleum HAC 896.00 372.50 36.74

CTL 93.19 38.74 3.82Yulin Excelle 3768.79 1264.12 125.78 CTM 1941.00 731.51 72.79 92.77

CTA 1098.23 413.89 41.18PVC 315.00 118.72 11.81

Future Energy 1678.36 1901.50 202.83 CTL 1527.31 1901.50 202.83 40.90(CV¼ 0.12, ε¼ 0.40, 1-

a¼ 0.95)5722.98 China Coal 4441.50 1646.32 161.40 CTM 2652.75 1068.79 104.78 96.56

MTO 784.00 315.87 30.97MTP 649.43 261.66 25.65



CTA 560.00 211.05 21.00PVC 325.12 122.53 12.19


a¼ 0.90)5578.91 China Coal 4081.09 1512.73 148.31 CTM 2652.75 1068.79 104.78 82.32

MTO 709.85 286.00 28.04MTP 392.00 157.94 15.48



CTA 560.00 211.05 21.00PVC 325.12 122.53 12.19


a¼ 0.80)5406.18 China Coal 3648.99 1352.56 132.60 CTM 2573.07 1036.69 101.64 67.61

MTO 392.00 157.94 15.48MTP 392.00 157.94 15.48



CTA 560.00 211.05 21.00PVC 325.12 122.53 12.19

Future Energy 1678.36 1901.50 202.83 CTL 1527.31 1901.50 202.83 40.90


water supply condition where CV¼ 0.18 and a park authority con-fidence level of 99%, water andwaste load allocation optimization isobtained, and a risk-averse park authority can reduce the economicefficiency losses to as low as possible. In this scenario, ε changesfrom 0.40 to a minimum constraint level with a step length of 0.05until no optimal solution exists. From Table 6, the economic effi-ciency losses can be controlled at 0.30, with the maximum totalpollution of 8112.53 tonnes being achieved when 1-a¼ 0.99 andCV¼ 0.18. It can be seen that a smaller ε leads to greater pollution inthe park; that is, the lower the economic efficiency loss expected bythe decision maker, the more difficult it is to control pollution.Therefore, the conflict between economic development and envi-ronmental protection are reflected mathematically in this scenario.The allocations for each company also have different sensitivities tothe changes in the economic efficiency loss tolerance levels.

4.3.1.5. Scenario 5: CV¼ 0.25, ε ¼ 0:40, 1-a is changing. In thisscenario, it is assumed that the park authority has an optimisticattitude towards the economic efficiency losses; that is ε ¼ 0:40,;and the water supply varies greatly; that is CV¼ 0.25. A set of so-lutions are obtained from the change of confidence intervals, asshown in Table 7. Comparing the results at the different confidencelevels, it can be seen that the total pollution increases noticeably

with an increase in 1-a. When the park authority has the highestrisk aversion degree, the total pollution damage to the environmentis 7072.08 tonnes, a rise of 1258.68 tonnes than at a confidencelevel of 80%, which proves that the attitude towards economic riskhas a significant impact on water and waste load allocation stra-tegies and the realization of environmental goals. The allocationresults for each company indicate that not all companies are sen-sitive to the changes in 1-a. The China Coal and Future Energycompanies, for example, are allocated a certain amount of waterand waste load regardless of the changes in the confidence level;however, the Yulin Excelle company has increasing resource allo-cations and the Yanchang Petroleum company receives a suddenwater and waste load resource increase when 1-a¼ 0.99. Inconclusion, the sensitivity reactions at the four companies underthe different resource consumption, pollutant emissions and pro-duction plans are not in complete accord.

4.3.1.6. Scenario 6: CV¼ 0.25, 1-a¼ 0.99, ε is changing. Under a veryunstable water supply condition at CV¼ 0.25, by setting 1-a¼ 0.99,the optimization model is run to assess the sensitivity of the allo-cations strategy to a changing ε. In this scenario, with a changeinterval of 0.05, ε reduces from 0.40 until no optimal solution exists.From the results shown in Table 8, the economic efficiency losses

Table 4Optimal results under different tolerance levels for economic return losses with CV¼ 0.12 and 1-a¼ 0.99.




Allocated loads Allocatedto

Allocation results Total costs (108

RMB)

COD(tonne)

NH3-N(tonne)

projects Water(104m3)

COD load(tonne)

NH3-N load(tonne)

(CV¼ 0.12, 1-a¼ 0.99,ε¼ 0.35)




CTA 560.00 211.05 21.00PVC 315.00 118.72 11.81

Future Energy 1678.36 1901.50 202.83 CTL 1527.31 1901.50 202.83 40.90(CV¼ 0.12, 1-a¼ 0.99,

ε¼ 0.30)6745.03 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 997.36 375.88 37.40PVC 315.00 118.72 11.81


ε¼ 0.25)7139.20 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63


ε¼ 0.20)7551.36 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63


ε¼ 0.15)8611.78 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63



can be controlled at 0.30 when CV¼ 0.25 and 1-a¼ 0.99, resultingin a total pollution load of 9006.49 tonnes. The allocation results foreach company indicate that the China Coal and Yulin Excellecompanies are allocated insensitive water and waste loads withchanges in ε, and the Yanchang Petroleum and the Future Energycompanies have different sensitivities to the changes in ε.

4.3.2. Comprehensive discussionBased on the scenario analysis, a detailed discussion is given in

this section.

Proposition 1. An equilibrium strategy is able to trade off theenvironmental and economic objectives

An IE approach gives guidance on reducing environmentaldamage while improving or sustaining economic performance

(Alper, 1992). From the scenario results, total pollution and eco-nomic efficiency losses can be balanced with an emissions reduc-tion. When the economic efficiency losses are below 15%, the totalpollution produced by the park industrial companies can becontrolled at 8611.78 tonnes. When ε increase, the total pollutionshows a downward trend to a trough of 5966.33 tonnes whenCV¼0.12 and the confidence level is set at 0.99. In other words, ifthe economic efficiency is raised, there is a commensurate increasein pollutant emissions even if there are emissions reductions andwater quality constraints. This result mathematically reflects theconflicts between the economic and environmental objectives inEqs. (1) and (2). As shown in Tables 4, 6 and 8, integrating water useand pollution control can find equilibrium water resource andwaste load strategies to achieve balanced efficient economic per-formance and control pollution damage to the environment under

Table 5Optimal results under different confidence levels with CV¼ 0.18 and ε¼ 0.40.






RMB)

COD(tonne)

NH3-N(tonne)


COD load(tonne)

NH3-N load(tonne)

(CV¼ 0.18, ε¼ 0.40, 1-a¼ 0.99)




CTA 1120.00 422.10 42.00PVC 408.89 154.10 15.33


a¼ 0.95)6324.38 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 560.00 211.05 21.00PVC 325.12 118.72 11.81


a¼ 0.90)6053.06 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 560.00 211.05 21.00PVC 315.00 118.72 11.81


a¼ 0.80)5772.15 China Coal 4451.92 1650.18 161.78 CTM 2652.75 1068.79 104.78 96.97

MTO 784.00 315.87 30.97MTP 659.02 265.52 26.03



CTA 560.00 211.05 21.00PVC 325.12 122.53 12.19








RMB)

COD(tonne)

NH3-N(tonne)


COD load(tonne)

NH3-N load(tonne)

(CV¼ 0.18, 1-a¼ 0.99,ε¼ 0.35)




CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63


ε¼ 0.30)8112.53 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63




a certain scenario. In conclusion, the equilibrium strategies devel-oped using the proposed model could help achieve the coordinatedaims of reducing environmental stress and improving resourceefficiency, as advocated by IE (Karn and Bauer, 2001).

Proposition 2. The application of IE encourages water resourcerecycling

Based on the IE principles, optimal water and waste load allo-cations were obtained under different scenarios for the four com-panies, as shown in Fig. 3. In scenario 1, 3, and 5, the China Coal andYulin Excelle companies gained greater water resource allocationswhen the confidence interval increased. At the same time, theFuture Energy company allocation was insensitive to changes in 1-a. The Yanchang Petroleum company was allocated 1859.31 104 ofwater under the scenarios except when 1-a¼0.99 in scenario 5,when the highest water allocation of 2236.74 104 was given. achanging allocation trend can be seen, in that three companiesexcept for the China Coal company are given a greater water allo-cation when the park authority expects a decrease in the economicefficiency loss risk. As ε reduces from 0.40 to a lower level with astep length of 0.05, the water allocated to all coal chemical com-panies increases under all examined water supply variation

Table 7Optimal results under various confidence levels with CV¼ 0.25 and ε¼ 0.40.




Allocat

COD(tonne

(CV¼ 0.25, ε¼ 0.40, 1-a¼ 0.99)

7072.08 China Coal 4587.77 1700.5

Yanchang 2236.74 818.32Petroleum

Yulin Excelle 6328.09 2122.5

Future Energy 1678.36 1901.5(CV¼ 0.25, ε¼ 0.40, 1-

a¼ 0.95)6486.37 China Coal 4587.77 1700.5




a¼ 0.90)6181.09 China Coal 4587.77 1700.5




a¼ 0.80)5813.40 China Coal 4587.77 1700.5

YanchangPetroleum

1859.31 680.23


Future Energy 1678.36 1901.5

conditions. Therefore, from the scenario analyses, it can beconcluded that when the park authority has a higher confidenceand risk tolerance so that 1-a increases and ε decreases, the totalwater andwaste load allocations increase because of the increase ininternal water reuse rather than external water inflow, clearlydemonstrating that the proposed method in this paper canencourage greater water recycling and increase overall park sus-tainability (Jia et al., 2016).

Proposition 3. Changes in the external environment affectobjective realization and allocation strategies

Changes in rainfall impact the quality and quantity of the watersupply. When the external environment changes, total pollutionfluctuates and the economic efficiency losses can be controlled atdifferent levels. The results from Supplementary Material showshow the environmental objective reflects the changes in CV, fromwhich it can be seen that total pollution is sensitive to changes inwater supply variability. In scenario 1, the total pollution is 5966.33tonnes when the park authority has a high degree of confidence inthe economic efficiency losses. When the water supply variationincreases to 0.25, the total pollution grows by 18.53%, indicatingthat uncertain water resources makes it more difficult to control

ed loads Allocatedto


RMB)

)NH3-N(tonne)


COD load(tonne)

NH3-N load(tonne)

4 166.72 CTM 2652.75 1068.79 104.78 102.34MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97

80.70 CTM 647.00 268.98 26.53 44.24HAC 1232.34 512.33 50.53CTL 89.00 37.00 3.65

6 211.20 CTM 3882.00 1463.03 145.58 133.95CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63

0 202.83 CTL 1527.31 1901.50 202.83 40.904 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97

67.08 CTM 647.00 268.98 26.53 37.01HAC 896.00 372.50 36.74CTL 93.19 38.74 3.82

3 170.95 CTM 3683.63 1388.27 138.14 87.90CTA 560.00 211.05 21.00PVC 315.00 118.72 11.81

0 202.83 CTL 1527.31 1901.50 202.83 40.904 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97

67.08 CTM 647.00 268.98 26.53 37.01HAC 896.00 372.50 36.74CTL 93.19 38.74 3.82

7 142.81 CTM 2933.33 1105.50 110.00 78.89CTA 560.00 211.05 21.00PVC 315.00 118.72 11.81

0 202.83 CTL 1527.31 1901.50 202.83 40.904 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97

67.08 CTM 647.00 268.98 26.53 37.01

HAC 896.00 372.50 36.74CTL 93.19 38.74 3.82

9 108.92 CTM 1941.00 731.51 72.79 71.71CTA 560.00 211.05 21.00PVC 403.64 152.12 15.14

0 202.83 CTL 1527.31 1901.50 202.83 40.90







RMB)

COD(tonne)

NH3-N(tonne)


COD load(tonne)

NH3-N load(tonne)

(CV¼ 0.25, 1-a¼ 0.99,ε¼ 0.35)




CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63


ε¼ 0.30)9006.49 China Coal 4587.77 1700.54 166.72 CTM 2652.75 1068.79 104.78 102.34

MTO 784.00 315.87 30.97MTP 784.00 315.87 30.97



CTA 1120.00 422.10 42.00PVC 630.00 237.43 23.63



total pollution. As shown in Tables 4, 6 and 8, the CVaR varies;under a lower water supply condition, the CVaR can be controlledwithin 0.15 to 0.40 with a change interval of 0.05; however, whenthe economic efficiency loss risk constraint reduces to a step lengthof 0.05, no optimal results can be achieved under medium or highwater supply variation levels. Unstable water supply conditions notonly influence the implementation of the total pollution require-ment based on IE theory, but also exacerbate the risk of economicefficiency losses. From a comparison of the allocation results inscenarios 3, 5, 7, it can be seen that the water resource and wasteload allocations at each company are sensitive to changes in CV; thespecific sensitivities are shown in Fig. 3. Therefore, adapting in-dustrial production to climate change affects resource and envi-ronmental management, which in turn leads to changes inindustrial activities (Martin and Watson, 2016).

Proposition 4. Coal chemical companies are sensitive at varyingdegrees towards changes in the decision making environment

The allocation results in the different scenarios in Fig. 4,indicatethe varying sensitivities at each company under different scenarios;as can be seen, each of the four companies have different sensi-tivities to changes in a parameter in the same scenario. Forexample, in scenario 1, the China Coal, Yanchang Petroleum andYulin Excelle companies developed their coal chemical projectproduction plans based on the water resource allocations from thepark authority, which fluctuate as 1-a changes. However, the FutureEnergy company is able to develop their CTL production plans witha guaranteed water resource of 1527.31104m3. In scenario 1, 3, and5, the resource allocations increase with the growth in 1-a. Inscenario 2, the Yulin Excelle company allocation increases; how-ever, in scenarios 4 and 6, as ε decreases, the allocation does notchange. From Fig. 5, the total costs (operating costs and environ-mental costs) partly change in the different scenarios. When thepark authority has a higher confidence degree, in scenarios 1, 3, and5, total costs increase for some companies but remain stable forothers. Therefore, it can be concluded that when the park authorityhas tighter economic efficiency loss risk control, where ε is setlower and 1-a is set close to 1, this could lead to an increase in costsas the environmental costs grow. In summary, the coal chemicalprojects are planned based on the water resource allocations from

the park authority and emitted pollutants to ensure water qualitystandards. Production plans based on the water resource andpollution emissions allocation were shown to be sensitive indifferent scenarios and had varying sensitivities to the changes,which was similar to the Knapsack problem (Sinha and Zoltners,1979).

4.3.3. Management proposalsBased on the computation results and discussion for the Yuheng

coal chemical park, some management proposals are given.

(1) Ecological concepts should be integrated to guide sustainablesystem development

Joseph Alper stated that “Industrial ecology enable industry toimprove the efficiency of its process” (Alper, 1992). The applicationof IE principles and methods can promote the coordinated devel-opment of the economy and the environment (Karn and Bauer,2001). Through the coupling of the industrial economic systemand the ecological system, total water pollution and economic ef-ficiency are highlighted as the major objectives in the proposedmodel in Eq. (5). As water quality and water pollutant reduction areset as the constraints, the proposedmethodology is able to combinethe controlled emission reductions while keeping the otherpollutant emissions within environmental regulations. As dis-cussed in Proposition 1 and Proposition 2, the integration of coaland water resource use and environmental protection through thestrategic planning in the proposed methodology can achieve abalance between environmental protection and economic perfor-mance. Therefore, the IE principles should be applied to addresssustainability and give insight into possible solutions.

(2) Coordinate conflicts through systematic management sys-tem collaboration

The IE has important principles for industry reformation that aregradually being embraced by leaders and policy-makers to addressthe inherent conflicts between economic growth and ecologicalenvironmental protection (Pauliuk et al., 2017). As the proposedmethodology simultaneously considers the economic and

Fig. 3. Water allocation results under different scenarios.

Fig. 4. Resource allocations to coal chemical projects under different scenarios.


environmental effects, the park authority and coal chemical com-panies can work together to achieve greater benefits than workingalone. According to the results and the discussion in Proposition 3,there are trade-offs in this Stackelberg game between the parkauthority goals and the coal chemical stockholder' objectives.

Therefore, it is more productive to integrate the skills and knowl-edge of the various decision makers to jointly resolve the conflicts.A system based on IE and the Stackelberg game allows for theclarification and reassignment of environmental responsibilities(for example, between the authority, the companies and the

Fig. 5. Costs to each company under different scenarios (CNY). Note: the total costs and operating costs are on the left vertical axis; the environment costs are on the right verticalaxis.


production projects), thereby reducing conflicts and promoting thedesign of equilibrium development strategies. There is an urgentneed to establish a collaborative systematic system to coordinatethe conflicts.

(3) Develop sensitive adaptation strategies that respond toenvironmental changes

As there are future uncertainties because of the inherent

variabilities in the ecosystem, scenarios should be used to explorethe potential consequences of the different response strategies (Xuand Li, 2012; Martin andWatson, 2016). As discussed in Proposition3, he larger the variation in available water, the more the emissionsand the larger the economic efficiency losses. Coal chemical pro-duction plans based on water and waste load allocations are sen-sitive to changes in the water environment, as discussed inProposition 4. Changes in rainfall impact water resources with re-gard to quality and quantity, can reduce accessibility to suitable


water supplies, and increase water quality management problems.The scenario results conducted in subsection 4.3.1 provide valuablesupport for decision makers so they can adjust their practices tocope with climate change. The application of scenario analysescould address different uncertainties and help establish an effectivelink between possible climate change effects and decision makerresponses; therefore, to guarantee sustainable development, it isnecessary to develop sensitive adaptation strategies for possiblefuture changes.

5. Conclusion and future study

In this study, a systematic decision making framework based onIE was developed to support efficient water utilization and pollu-tion control in coal chemical industrial parks. In the proposedmethodology, pollution damage and economic efficiency weresystematically coordinated and the conflicting objectives of thepark authority and the coal chemical companies on two levels wereeffectively quantified in trade-offs through a leader-followerStackelberg game decision process. The results and discussionfrom the case study demonstrated that the economic efficiencyperformance and environmental pollutant damage can be balancedby equilibriumwater resource and waste load allocation strategies.The application of IE has the potential to encourage water recyclingin the park to reduce the risk of economic efficiency losses. Byconstructing a set of scenarios that measured the effects of theactual water supply conditions and the different decision maker'attitudes, sensitivity analyses were conducted and sensitive adap-tation strategies provided to respond to various environmental andpolicy changes. With such a solution-based approach, the proposedmodel could be easily applied to industrial parks or zones facingwater resource shortages and pollution control problems. Thesepropositions and management proposals combined with the sce-narios can assist decision makers involved in bi-level sequentialdecisions make adjustments to their control targets to make moreinformed decisions for overall sustainable IE development. Infuture research, the model could be improved to incorporate en-ergy and water resource use to develop models for industrial eco-systems that include partnership exchanges inside industries orwith other industries, and could also be extended to a systemicsolution-driven approach towards sustainability that mitigates theeffects of climate change.

Acknowledgements

This work is supported by State Key Development Program of(for) Basic Research of China [973 Program, Grant No.2011CB201200], the National Natural Science Foundation of China[Grant No. 71771157, 71301109], the Funds for Creative ResearchGroups of China [Grant No. 50221402], Soft Science Program ofSichuan Province [Grant No. 2017ZR0154], and Funding of SichuanUniversity [Grant No. skqx201726], China Postdoctoral ScienceFoundation Funded Project [Grant No. 184089], and the Funda-mental Research Funds for the Central Universities[2012017yjsy104].

Appendix A. Supplementary data

Supplementary data related to this article can be found athttps://doi.org/10.1016/j.jenvman.2018.04.080.

References

Alper, J., 1992. Introducing industrial ecology. Nat. Biotechnol. 10, 1522e1522.Anandalingam, G., Apprey, V., 1991. Multi-level programming and conflict

resolution. Eur. J. Oper. Res. 51, 233e247.Boix, M., Montastruc, L., Azzaro-Pantel, C., Domenech, S., 2015. Optimization

methods applied to the design of eco-industrial parks: a literature review.J. Clean. Prod. 87, 303e317.

China's National Energy Administration, 2015. Action Plan for Clean and EfficientUse of Coal (2015e2020). http://zfxxgk.nea.gov.cn/auto85/201505/t20150505_1917.htm.

Chu, S., Majumdar, A., 2012. Opportunities and challenges for a sustainable energyfuture. Nature 488, 294.

Edgington, S.M., 1995. Industrial ecology: biotech's role in sustainable development.Nat. Biotechnol. 13, 31e34.

Grafton, R.Q., Pittock, J., Davis, R., Williams, J., Fu, G., Warburton, M., Udall, B.,Mckenzie, R., Yu, X., Che, N., 2013. Global insights into water resources, climatechange and governance. Nat. Clim. Change 3, 315e321.

Griggs, D., StaffordeSmith, M., Gaffney, O., Rockstr�om, J., Ohman, M.C.,Shyamsundar, P., Steffen, W., Glaser, G., Kanie, N., Noble, I., 2013. Sustainabledevelopment goals for people and planet. Nature 495, 305e307.

Han, L., Song, Y., Duan, L., Yuan, P., 2015. Risk assessment methodology for ShenyangChemical Industrial Park based on fuzzy comprehensive evaluation. Environ.Earth Sci. 73, 5185e5192.

Hu, Z., Wei, C., Yao, L., Li, L., Li, C., 2016. A multi-objective optimization model withConditional Value-at-Risk constraints for water allocation equality. J. Hydrol.542, 330e342.

Jia, S., Zhuang, H., Han, H., Wang, F., 2016. Application of industrial ecology in waterutilization of coal chemical industry: a case study in Erdos, China. J. Clean. Prod.135, 20e29.

Karn, B.P., Bauer, D., 2001. Sustainability: insight from industrial ecology. Science293, 1995.

Kavouridis, K., Koukouzas, N., 2008. Coal and sustainable energy supply challengesand barriers. Energy Policy 36, 693e703.

Kong, X.M., Huang, G.H., Fan, Y.R., Li, Y.P., Zeng, X.T., Zhu, Y., 2017. Risk analysis forwater resources management under dual uncertainties through factorial anal-ysis and fuzzy random value-at-risk. Stoch. Environ. Res. Risk Assess. 1e16.

Leong, Y.T., Lee, J.Y., Tan, R.R., Ji, J.F., Chew, I.M.L., 2017. Multi-objective optimizationfor resource network synthesis in eco-industrial parks using an integratedanalytic hierarchy process. J. Clean. Prod. 143, 1268e1283.

Li, J., Hu, S., 2017. History and future of the coal and coal chemical industry in China.Resour. Conserv. Recycl. 124, 13e24.

Lin, B., Long, H., 2015. A stochastic frontier analysis of energy efficiency of China'schemical industry. J. Clean. Prod. 87, 235e244.

Liu, B., 1998. Stackelberg-nash equilibrium for multilevel programming with mul-tiple followers using genetic algorithms. Comput. Math. Appl. 36, 79e89.

Liu, J., Mooney, H., Hull, V., Davis, S.J., Gaskell, J., Hertel, T., Lubchenco, J., Seto, K.C.,Gleick, P., Kremen, C., 2015. Sustainability. systems integration for global sus-tainability. Science 347, 1258832.

Lovelady, E.M., El-Halwagi, M.M., 2009. Design and integration of eco-industrialparks for managing water resources. Environ. Prog. Sustain. Energy 28,265C272.

Man, Y., Xiao, H., Cai, W., Yang, S., 2017. Multi-scale sustainability assessments forbiomass-based and coal-based fuels in China. Sci. Total Environ. 599e600, 863.

Martin, T.G., Watson, J.E.M., 2016. Intact ecosystems provide best defence againstclimate change. Nat. Clim. Change 6, 122e124.

Miller, A.C., Rice, T.R., 1983. Discrete approximations of probability distributions.Manag. Sci. 29, 352e362.

Moghaddam, I.G., Nick, M., Fallahi, F., Sanei, M., Mortazavi, S., 2013. Risk-averseprofit-based optimal operation strategy of a combined wind farm cascadehydro system in an electricity market. Renew. Energy 55, 252e259.

Nash, J., 1951. Non-cooperative games. Ann. Math. 286e295.Niu, D.X., Song, Z.Y., Xiao, X.L., 2017. Electric power substitution for coal in China:

status quo and SWOT analysis. Renew. Sustain. Energy Rev. 70, 610e622.Pauliuk, S., Arvesen, A., Stadler, K., Hertwich, E.G., 2017. Industrial ecology in inte-

grated assessment models. Nat. Clim. Change 7, 13e20.People, C.N., 2008. Environmental protection Law of the People's republic of china.Poff, L.R., Brown, C.M., Grantham, T.E., Matthews, J.H., Palmer, M.A., Spence, C.M.,

Wilby, R.L., Haasnoot, M., Mendoza, G.F., Dominique, K.C., 2016. Sustainablewater management under future uncertainty with eco-engineering decisionscaling. Nat. Clim. Change 6 (1), 25.

Qi, L., Jiang, H., 1997. Semismooth Karush-Kuhn-Tucker equations and convergenceanalysis of Newton and quasi-Newton methods for solving these equations.Math. Oper. Res. 22, 301e325.

Ren, J., Liang, H., Liang, D., Sun, L., Gao, Z., 2016. Design for sustainability of in-dustrial symbiosis based on emergy and multi-objective particle swarm opti-mization. Sci. Total Environ. 562, 789e801.

Rockafellar, R.T., Uryasev, S., 2010. Optimization of conditional value-at-risk. J. Risk29, 1071C1074.

Rubiocastro, E., Ponceortega, J.M., N¢polesrivera, F., Elhalwagi, M.M., Sernagonz¢lez, M., Jimnezgutirrez, A., 2010. Water integration of eco-industrial parks usinga global optimization approach. Ind. Eng. Chem. Res. 49, 9945e9960.

Saharidis, G.K., Ierapetritou, M.G., 2009. Resolution method for mixed integer bi-level linear problems based on decomposition technique. J. Glob. Optim. 44,29e51.

Service, R.F., 2017. Cleaning up coal-cost-effectively. Science 356.Shen, B., Han, Y., Price, L., Lu, H., Liu, M., 2017. Techno-economic evaluation of

strategies for addressing energy and environmental challenges of industrialboilers in China. Energy 118, 526e533.

https://doi.org/10.1016/j.jenvman.2018.04.080http://refhub.elsevier.com/S0301-4797(18)30462-6/sref1http://refhub.elsevier.com/S0301-4797(18)30462-6/sref1http://refhub.elsevier.com/S0301-4797(18)30462-6/sref2http://refhub.elsevier.com/S0301-4797(18)30462-6/sref2http://refhub.elsevier.com/S0301-4797(18)30462-6/sref2http://refhub.elsevier.com/S0301-4797(18)30462-6/sref3http://refhub.elsevier.com/S0301-4797(18)30462-6/sref3http://refhub.elsevier.com/S0301-4797(18)30462-6/sref3http://refhub.elsevier.com/S0301-4797(18)30462-6/sref3http://zfxxgk.nea.gov.cn/auto85/201505/t20150505_1917.htmhttp://zfxxgk.nea.gov.cn/auto85/201505/t20150505_1917.htmhttp://refhub.elsevier.com/S0301-4797(18)30462-6/sref5http://refhub.elsevier.com/S0301-4797(18)30462-6/sref5http://refhub.elsevier.com/S0301-4797(18)30462-6/sref6http://refhub.elsevier.com/S0301-4797(18)30462-6/sref6http://refhub.elsevier.com/S0301-4797(18)30462-6/sref6http://refhub.elsevier.com/S0301-4797(18)30462-6/sref7http://refhub.elsevier.com/S0301-4797(18)30462-6/sref7http://refhub.elsevier.com/S0301-4797(18)30462-6/sref7http://refhub.elsevier.com/S0301-4797(18)30462-6/sref7http://refhub.elsevier.com/S0301-4797(18)30462-6/sref8http://refhub.elsevier.com/S0301-4797(18)30462-6/sref8http://refhub.elsevier.com/S0301-4797(18)30462-6/sref8http://refhub.elsevier.com/S0301-4797(18)30462-6/sref8http://refhub.elsevier.com/S0301-4797(18)30462-6/sref8http://refhub.elsevier.com/S0301-4797(18)30462-6/sref8http://refhub.elsevier.com/S0301-4797(18)30462-6/sref9http://refhub.elsevier.com/S0301-4797(18)30462-6/sref9http://refhub.elsevier.com/S0301-4797(18)30462-6/sref9http://refhub.elsevier.com/S0301-4797(18)30462-6/sref9http://refhub.elsevier.com/S0301-4797(18)30462-6/sref10http://refhub.elsevier.com/S0301-4797(18)30462-6/sref10http://refhub.elsevier.com/S0301-4797(18)30462-6/sref10http://refhub.elsevier.com/S0301-4797(18)30462-6/sref10http://refhub.elsevier.com/S0301-4797(18)30462-6/sref11http://refhub.elsevier.com/S0301-4797(18)30462-6/sref11http://refhub.elsevier.com/S0301-4797(18)30462-6/sref11http://refhub.elsevier.com/S0301-4797(18)30462-6/sref11http://refhub.elsevier.com/S0301-4797(18)30462-6/sref12http://refhub.elsevier.com/S0301-4797(18)30462-6/sref12http://refhub.elsevier.com/S0301-4797(18)30462-6/sref13http://refhub.elsevier.com/S0301-4797(18)30462-6/sref13http://refhub.elsevier.com/S0301-4797(18)30462-6/sref13http://refhub.elsevier.com/S0301-4797(18)30462-6/sref14http://refhub.elsevier.com/S0301-4797(18)30462-6/sref14http://refhub.elsevier.com/S0301-4797(18)30462-6/sref14http://refhub.elsevier.com/S0301-4797(18)30462-6/sref14http://refhub.elsevier.com/S0301-4797(18)30462-6/sref15http://refhub.elsevier.com/S0301-4797(18)30462-6/sref15http://refhub.elsevier.com/S0301-4797(18)30462-6/sref15http://refhub.elsevier.com/S0301-4797(18)30462-6/sref15http://refhub.elsevier.com/S0301-4797(18)30462-6/sref16http://refhub.elsevier.com/S0301-4797(18)30462-6/sref16http://refhub.elsevier.com/S0301-4797(18)30462-6/sref16http://refhub.elsevier.com/S0301-4797(18)30462-6/sref17http://refhub.elsevier.com/S0301-4797(18)30462-6/sref17http://refhub.elsevier.com/S0301-4797(18)30462-6/sref17http://refhub.elsevier.com/S0301-4797(18)30462-6/sref18http://refhub.elsevier.com/S0301-4797(18)30462-6/sref18http://refhub.elsevier.com/S0301-4797(18)30462-6/sref18http://refhub.elsevier.com/S0301-4797(18)30462-6/sref19http://refhub.elsevier.com/S0301-4797(18)30462-6/sref19http://refhub.elsevier.com/S0301-4797(18)30462-6/sref19http://refhub.elsevier.com/S0301-4797(18)30462-6/sref20http://refhub.elsevier.com/S0301-4797(18)30462-6/sref20http://refhub.elsevier.com/S0301-4797(18)30462-6/sref20http://refhub.elsevier.com/S0301-4797(18)30462-6/sref21http://refhub.elsevier.com/S0301-4797(18)30462-6/sref21http://refhub.elsevier.com/S0301-4797(18)30462-6/sref21http://refhub.elsevier.com/S0301-4797(18)30462-6/sref22http://refhub.elsevier.com/S0301-4797(18)30462-6/sref22http://refhub.elsevier.com/S0301-4797(18)30462-6/sref22http://refhub.elsevier.com/S0301-4797(18)30462-6/sref23http://refhub.elsevier.com/S0301-4797(18)30462-6/sref23http://refhub.elsevier.com/S0301-4797(18)30462-6/sref23http://refhub.elsevier.com/S0301-4797(18)30462-6/sref24http://refhub.elsevier.com/S0301-4797(18)30462-6/sref24http://refhub.elsevier.com/S0301-4797(18)30462-6/sref24http://refhub.elsevier.com/S0301-4797(18)30462-6/sref24http://refhub.elsevier.com/S0301-4797(18)30462-6/sref25http://refhub.elsevier.com/S0301-4797(18)30462-6/sref25http://refhub.elsevier.com/S0301-4797(18)30462-6/sref26http://refhub.elsevier.com/S0301-4797(18)30462-6/sref26http://refhub.elsevier.com/S0301-4797(18)30462-6/sref26http://refhub.elsevier.com/S0301-4797(18)30462-6/sref27http://refhub.elsevier.com/S0301-4797(18)30462-6/sref27http://refhub.elsevier.com/S0301-4797(18)30462-6/sref27http://refhub.elsevier.com/S0301-4797(18)30462-6/sref28http://refhub.elsevier.com/S0301-4797(18)30462-6/sref29http://refhub.elsevier.com/S0301-4797(18)30462-6/sref29http://refhub.elsevier.com/S0301-4797(18)30462-6/sref29http://refhub.elsevier.com/S0301-4797(18)30462-6/sref29http://refhub.

Documents

Equilibrium approach towards water resource management ......water directly intothe river (Kavouridis and Koukouzas, 2008). The situation of severe water shortages, water pollution,